Yiming Zeng

Papers

Total Citations

H-Index

About

Yiming Zeng is an emerging robotics researcher whose work sits at the intersection of generative AI and autonomous robot learning. His research primarily focuses on applying diffusion-based models to core robotics challenges, including visual navigation and functional object rearrangement — two domains critical to developing intelligent, adaptable robotic systems. Zeng's most notable contribution, "LVDiffusor" (2024), has garnered 7 citations and demonstrates his innovative approach of distilling functional rearrangement priors from large pretrained models into diffusion frameworks, enabling robots to handle diverse objects and configurations with greater versatility. His follow-up work, "NaviDiffusor" (2025), extends this philosophy to mobile robotics, introducing a cost-guided diffusion model that bridges the gap between classical geometric navigation methods and modern learning-based approaches, earning 3 citations shortly after publication. His broader research agenda challenges conventional assumptions in robot learning — particularly how prior knowledge can be embedded into generative models to improve both adaptability and reliability. With a growing citation record and a consistent focus on practical, high-impact applications, Zeng represents a promising voice in next-generation embodied AI research, with work that is increasingly relevant to both academic and real-world robotics communities.